Studies of the efficacy of tricyclic antidepressants (TCA) in patients with endogenous depression regularly report that 20-40% of such patients are either intolerant of, or refractory to, TCA treatment. Monoamine oxidase inhibitor antidepressants (MAOI) are infrequently used in these patients because of the belief among clinicians that MAOIs are not effective in endogenous depression. This is so despite a variety of studies, including our own pilot work, which suggest that MAOIs may be effective in this group. Available controlled studies do not adequately address the question, mainly because of methodologic shortcomings by current standards of clinical trials. Recent controlled studies have shown a preferential response to MAOIs over TCAs among patients with atypical depression, confirming previous clinical observations. In contrast to the mood unresponsivity, anorexia, and insomnia characteristic of endogenous depression, patients with atypical depression show mood response to favorable events, overeating, and oversleeping. This has raised the important issue of whether MAOI superiority to TCA is confined to atypical depression or may be seen in other subgroups of major depression. We propose to address the clinical and theoretical issues raised above by conducting a double blind, controlled trial of phenolzine compared to imipramine and placebo in 120 outpatients with melancholia and a trial of phenelzine compared to imipramine in 60 inpatients with melancholia. This is to be a collaborative project between the New York State Psychiatric Institute and Hillside Hospital to have an adequate sample available to study. (The portion of the work to be done at Long Island Jewish-Hillside Medical Center will be described in a separate grant.) The trial would provide data on the utility of MAOIs in endogenous depression, clarifying the relative responses to MAOIs in atypical and endogenous depression as well as diminishing the current obstacles to the use of MAOIs in endogenous patients unresponsive to TCAs. If MAOIs are shown to be effective in melancholia, this would provide empirical support for further trials in refractory patients and provide the impetus to develop new MAOIs with fewer side effects.